Description of problem:
Virtual Machine Management Guide:
6.8.1. Creating a Snapshot of a Virtual Machine
The 'IMPORTANT' note is wrong. I.e. no need to manually freeze and thaw the guest filesystem when using the REST API. It is done automatically by the engine. These actions where added just as a safety measure to mitigate issues with live snapshotting a VM with Cinder disks. E.g. it should be invoked manually if the VM stayed on freeze state after taking a live snapshot (due to a failure).
Procedure 6.10. Creating a Snapshot of a Virtual Machine
The note is wrong. 'thaw' should be manually executed only in case of a failure (only if the VM stayed on freeze state).
Why is this required for Cinder? Don't we do it for these VMs?
(In reply to Yaniv Dary from comment #1)
> Why is this required for Cinder? Don't we do it for these VMs?
We do it automatically, i.e. the special notes for the rest-api in the documentation are redundant.
(In reply to Daniel Erez from comment #2)
> (In reply to Yaniv Dary from comment #1)
> > Why is this required for Cinder? Don't we do it for these VMs?
> We do it automatically, i.e. the special notes for the rest-api in the
> documentation are redundant.
I still see this issue in the 4.0 docs:
Can we please fix this?
Assigning to Byron for review.
This content has been published to: